1. Field of the Invention
The invention relates to a semiconductor substrate, comprising a carrier wafer and a layer of single-crystal semiconductor material, to a process for producing the semiconductor substrate, and to an intermediate product formed in the process.
2. Background Art
Wafers comprising a thin layer of semiconductor material on an electrically insulating substrate have been disclosed as preliminary products for the fabrication of electronic components. In an alternative configuration, the thin semiconductor layer may be separated from a substrate which likewise consists of semiconductor material, by an electrically insulating layer. If the semiconductor material of the thin layer is silicon, the wafers are known as SOI (silicon on insulator) wafers.
A number of processes for producing wafers of this type are also known. In most of the known processes, a separating layer, for example a layer with cavities, is produced just below the surface of a semiconductor wafer (known as the donor wafer). The donor wafer which has been prepared in this manner is joined to a second wafer, the carrier wafer. Then, the donor wafer is split along the separating layer, transfering a layer of the donor wafer to the carrier wafer.
WO03/003430A2 describes one process in which a thin layer of semiconductor material is transferred from a donor wafer to a carrier wafer. First, patterns of periodically recurring recesses of predetermined geometries are produced on that side of the donor wafer which is intended for the transfer. These recesses are then closed up at the surface of the donor wafer by a heat treatment, so as to form a layer with periodically recurring cavities beneath a continuous layer at the surface of the material. The donor wafer which has been prepared in this manner is joined to a carrier wafer. Then, the donor wafer is split along the layer containing the cavities, for example by further heat treatment.
The aforementioned process comprises a large number of steps and is therefore relatively complex. Furthermore, the process does not allow layers to be transferred which are as thin as desired, since the thickness of the layer is limited by the lithography used to produce the recesses. To obtain very thin layers, for example those with a thickness of less than 10 nm, it is necessary for a thicker layer, such as one with a thickness of 50 nm, first to be transferred to the carrier wafer, followed by reducing the layer thickness by suitable measures as described in WO03/003430 A2.
By way of example, it is possible for a layer with a mean thickness of 100 nm and a standard deviation of 5%, based on the mean layer thickness, to be transferred. This means that up to 32% of the surface area deviates by 5% (i.e. 5 nm) or more and even 0.3% of the surface area even deviates by 15% (i.e. 15 nm) or more from the mean layer thickness. If the thickness of the transferred layer is then reduced to 15 nm, the standard deviation of 5% which is present after transfer and separation leads to the transferred layer of semiconductor material being completely removed in statistical terms over approximately 0.15% of the surface area. In the case of a wafer with a diameter of 300 mm and a surface area of 707 cm2, therefore, the layer of semiconductor material is completely removed over a surface area of approximately 1 cm2. These regions are detectable as HF defects. If the thickness of a transferred semiconductor layer is excessively reduced in the manner described, the layer thickness homogeneity which is present after the transfer and separation has a direct effect on the HF defect density after the thickness reduction. In addition, the conventional processes used for thickness reduction tend to have an adverse effect on the absolute layer thickness homogeneity, and consequently at very low final thicknesses the HF defect density rises still further.